1

Tourism productivity: incentives and obstacles to fostering growth

Stefania Lionetti1

1 Istituto di Ricerche Economiche, University of Lugano. Tel +41 586664790; fax +41 586664662. I would like to thank Professor Maggi for his support and help, Dr. Oscar Gonzalez for his hints, Professor Barros for the useful comments and the participants at the 57th AIEST and 1st CIATE conferences. All the errors remain mine.

2

Abstract

This paper intends firstly to estimate tourism productivity in 208 countries in the years 1990,

1995, 2000 and 2004. Secondly, it analyzes if the differential of productivity across countries could

be due to some structural characteristics of the countries themselves. The study uses a stochastic

production frontier approach and a technical efficiency model to analyze the determinants of

efficiency across countries. Private capital and labour result to be more influential than public capital

on the number of arrivals. The results suggest that the tertiary school enrolment, the level of

communication technologies, the country openness to international trade all significantly contribute to

efficiency.

JEL Classification: D24, L83, O10, O40, O50

1. Introduction

International tourism represents the principal exports source for 83% of developing

countries. Tourism is considered as an important source of foreign exchange earnings,

employment of domestic labour and a source of growth for a country. But the tourism

industry is one of the least productive sectors in the most developed economies.

Sectors that suffer from below-average labour productivity tend to face procurement

problems in the factor markets. In fact tourism faces difficulties in both attracting capital

and attracting highly qualified staff (OECD 2006).

The paper takes into consideration two distinct traditions: Growth theory and stochastic

frontier models.

Endogenous growth theory suggests that economic growth is related to a) large scale

b) sectors with high intensity of R&D and high productivity. Moreover, growth theories

stress the role of human capital (Lucas 1988), public infrastructure (Barro 1990) or

incentives to innovate (Romer 1990).

3

More and more studies have been focusing on the relationship between tourism and

the economic growth rate; Pigliaru (1999) find that many small tourism countries have

grown faster compared to other countries. Lozano, Gomez, and Rey-Maquiera (2005)

state that the length of the growth period of a destination depends on the quality of

private tourism services; a stylized fact of this literature is that the fastest growth rate

countries are the ones with the highest tourism specialization but most of the literature

results demonstrate that tourism does not lead per se to economic growth (Neves

Sequeira and Campos 2005). The question of how relevant is the tourism sector for

the economic growth of an area is remarkable in a more and more services globalized

society. Eugenio-Martín, Martín Morales and Scarpa (2004) study how relevant tourism

sector is for the economic growth of the regions and vice versa. They show that the

tourism sector is significant for the economic growth of medium or low-income

countries, though not necessarily for developed countries.

The question posed by this study is about the relationship between tourism productivity

and the number of tourists arrivals.

Most of both theoretical and empirical studies in the economics of tourism focus on the

demand side. One of the contributions of this works is exploring some economics

empirics on the supply side of tourism. Tourism is not a homogeneous product. The

supply side combines the three main components: transport, tour operators and travel

agents, and accommodation. Tourism is a peculiar sector, also because it is a

collection of interrelated industries and thus comprises many services. Small and mid-

size businesses dominate in tourism, both in developing and developed countries (ILO

2001). In developing countries, in the hotel industry where international chains are

concerned, the labour market is organized in the same way as other industrial sectors,

while in other small-scale tourism activities things are different. Thus problems with the

definition and measurements of tourism productivity arise, such as the definition of the

output itself, the quality, and the impact of technical change.

4

Taking into account the difficulty of drawing conclusions about the relative productivity

in this industry without considering the mix and nature of the services provided

(Anderson et al. 1999), the literature in hospitality industry firstly used average

occupancy rates and average room/ rates as indicators of performance, labour cost

ratio; then there have been cost-volume-profit analysis. Nowadays the most used

techniques in analyzing productivity and efficiency belong to four groups: the Data

Envelopment Analysis (DEA), the Distributon-Free Technique, the Thick Frontier

Approach and the Stochastic Frontier Approach (Anderson, Fish, Xia and Michello

1999). All these techniques attempt to define variations from an efficient frontier using

alternative assumptions regarding the probability distributions of the X-efficiency and

random error. Among these techniques, the DEA assumes that there are no random

fluctuations from the efficient frontier model, thus considering all the deviations as

inefficiency. DEA, but is sensitive to the input/output specification of the model and can

only measure relative efficiency levels. DEA is very popular because it does not require

a functional form assumption. Its presence is strong in the literature, see Banker

(1993), Bessent and Bessent (1980), Brown and Ragsdale (2002), Chiang et al.

(2004), Cooper et al. (2004), Hwang and Chang (2003), Barros and Athanassiou

(2004), Barros and Alves (2004). The Distribution-Free Approach assumes that

efficiency differences are stable over time while random error will approach to zero. Is

limit is that requires data for many years in order for the random term to cancel out

(Berger, 1991, 1993; Berger and Humphrey, 1992a). The Thick Frontier Approach

divides the magnitude of the error terms into quartiles, and any differences in efficiency

within groups represents random error, while inefficiency differences between the

highest and lowest quartiles represent inefficiencies (Berger and Humphrey, 1991,

1992b; Berger 1993). The stochastic Frontier Approach not only provides overall

industry estimates but also individual firm estimates. Thus identifying units relatively

efficient, and determining the magnitude of firm’s inefficiencies we could suggests

5

paths in order to reduce these inefficiencies (Anderson, Fish, Xia and Michello 1999) .

Representative studies using traditional stochastic frontier techniques are to Ferrier

and Lovell (1990), Timme and Yang (1992), Simeone and Li (1997), and Anderson et

al. (1998b)

Usually the literature examining the efficiency of the hotel industry uses revenues as

proxy for output. Specifically, the total revenues are supposed to be generated from

rooms, gaming, food and beverage, and other (such as store spaces, laundry, beauty

salons, ball courts etc.). As far as the inputs are concerned the literature considers the

number of full- time equivalent employees, the number of rooms, total gaming related

expenses, total food and beverage expenses, and other operating expenses2. Morey

and Dittman (1995) find that the hotel industry operates at 89% efficiency. Efficiency is

often hard to evaluate because it is difficult to determine an efficient amount of

resources and generally, hotel organizations have not developed standard cost

estimates of outputs. That is why, differently from previous studies on tourism

productivity, in this study only real variables will be taken into consideration, and as a

proxy for output the number of international arrivals is taken into consideration.

This study intends firstly to estimate productivity in tourism from a production function

of tourism. Due to the availability of both data on private and public sectors capital

invested in tourism all over the world we could test the influence of both kinds of capital

on tourists international arrivals and the level of the efficiency of the hotel industry for

each country in the world. Owing to the lack of comparable data, at a country level I

could not use monetary variables as it have been done by previous studies.

Secondly, this study analyzes if the differential of productivity across countries could be

due to some structural characteristics of the countries themselves, thus verifying the

relationship between the institutional and governance structure and the economic

performance.

2 See Bell and Morey (1995), Anderson et a. (2000), Anderson, Fish, and Michello (1999), Barros (2004).

6

The study uses a stochastic production frontier approach in order to measure country

level technical efficiency for a better comparison of economic performances. Through

the stochastic production frontier model the paper also intends to seek for the

determinants of technical efficiency.

The results suggest that the percentage of tertiary school people, the level of

communication technology, and the country openness to international trade all

significantly contribute with the expected negative sign to inefficiency.

The remainder of the paper is organized as follows. The next section provides a

definition and a very brief overview on the debate on productivity and efficiency.

Section 3 explains the model and the methodologies applied. Section 4 shows the

results obtained. Section 5 concludes.

2. Productivity and efficiency when looking at growth

Productivity is the ratio of output to input. One firm is more productive than another if it

can produce more with the same inputs or it can produce the same output with less of

all inputs. The literature individualizes five drivers to productivity: investment,

innovation, skills, enterprise, and competition. The contention that these drivers are the

main ones of productivity has neither been theoretically nor empirically tested. But

regional development policy is very influenced by this set of indicators.

Investment in physical capital can affect economic growth both directly (because it

increases aggregate demand) and via augmented productivity. Investment can expand

the production possibility frontier without changing marginal products or increase the

marginal contribution of the capital. That is to say that, other things being constant,

economic growth could be due to a bigger stock of capital or to an introduction of new

capital of higher productivity. In the economic literature most of the studies (Bradford

Delong and Summers 1991, Jorgenson 2005) find a positive relationship between

investment in physical capital and economic growth, mostly based on the assumption

that successive vintages of capital embody technological progress.

7

Critics to this literature (Mankiw, Romer and Weil 1992, Blomstrom et al. 1996) state

that the causality relationship can go from growth to investment or that investment in

the long run has no effect on economic growth.

Innovation is a continuous process of discovery, learning and application of new

technologies and techniques (that are both cumulative and interdependent) from many

sources. Schumpeter (1934) was the first one to stress the importance of innovation to

fostering growth. Innovation includes both fundamental and applied innovation.

Moreover, it can also take the form of organizational changes and new market

strategies which expand demand for products, support existing structure for new

methods of production and increase the efficiency of the other types of innovative

effort, leading to productivity improvements. Nickell and Van Reenen (2001) show that

innovation can only impact productivity if it is spread around the economy.

Skills reflect the fact that more skilled labour is likely to be more productive. Two

strands of literature are relevant to this regard: 1) Lucas (1988) claiming that the

relationship between growth and productivity depends on the rate of accumulation of

human capital and 2) Nelson and Phelps (1966) assuming differently that the

relationship is due to the stock of human capital, thus implying that an increase in

human capital affects permanently economic growth.

Enterprise involves seizing new business opportunities (both for new and incumbent

firms) . The importance of a vibrant enterprise culture has long been recognized as

essential for growth. There are a number of features that contribute to the overall state

of enterprise in the economy: a) Entrepreneurship: the motivation for business creation

and business growth in an economy, seizing opportunities and being rewarded for

success. b) Socio-cultural attitudes: social and cultural norms influence a community’s

attitudes and preferences . c) Capital markets: efficient and effective financial markets

provide organizations and individuals with funding for new ventures and investments.

Competition: the effect of competition on productivity is not as clear-cut as the other

drivers of productivity. From one side competition forces prices to converge to marginal

costs and more competitive markets encourage cost-reducing improvements because

their higher price elasticity of demand means that there is more room for firms to

increase profits than in less competitive markets. From the other side, innovation thrive

in highly concentrated markets, because monopoly rents are eroded in competitive

ones. Aghion and Howitt (1998) tried to reconcile these two adverse effects stating that

the relationship between product market competition and productivity growth should be

positive in industries: a) characterised by weak control of managers by shareholders, b)

where tacit knowledge is the main limiting barrier to imitation relative to patent

protection, c) with low density of technologic-specific fixed investments.

In this paper we consider as the economic unit the country and not the firm, thus we

will construct a production function.

Productivity is used as a performance benchmark to rank firms and countries or as a

measure of the performance improvement over time. It is well known that productivity is

often associated with indicators of success. Nevertheless measuring productivity

carries ambiguity (Van Biesebroeck 2006). Assumptions on the production technology,

the functional form, the firm behaviour, the other unobservable elements that affect

output matter and can lead to different measures. The production function

relates inputs ( ) (it it ititQ A f X= ) itX to output . is an unobservable term that

varies across firms i and time periods t. There are different measures of productivity:

the ones mostly used are total factor productivity and labour productivity

itQ itA

3.

We could implement the same definitions at a country level. The idea here is to

measure the economic performance in the tourism sector in terms of arrivals at a

country level. Thus a country should manage its scarce inputs efficiently; and tourism

8

3 Given Yt = ZtF(Kt,Lt), Total Factor Productivity (TFP) is defined to be Yt/F(Kt,Lt). Labor productivity is the quantity of output per time spent or numbers employed.

9

destination performance can be evaluated through a measurement of efficiency

(Cracolici, Nijkamp, and Rietveld 2008).

Of course the arrivals variable include both a demand and a supply component, it is

basically a “demanded supply” variable4; taking this into mind, we then compare each

country’s output with the maximum output that could be obtained with the same

resources (that is to say the world best possible output), given

the available resources and technology in the period considered. We are thus

measuring technical efficiency.

The literature (e. g. Fare, Grosskopf et al. 1994) decomposes productivity into changes

in efficiency (catching up) and changes in technology (innovation). Each country is

compared to a frontier. How much a country is close to the world frontier stands for the

catch up (efficiency); if the world frontier moves, it means technical change or

innovation.

Measuring productive efficiency allows one to test competing hypotheses regarding

sources of efficiency or differentials in productivity. Moreover, such measurement

enables us to quantify the potential increases in output that might be associated with

an increase in efficiency (Farrel 1957). The economic literature on production efficiency

typically distinguishes two types of efficiency: technical efficiency and allocative

efficiency. The latter includes as components cost minimization, revenue maximization,

and profit maximization. A technically efficient implies producing the maximum output

for a given amount of inputs, conditional on the production technology available to it.

An allocative efficient “economic unit” applies the optimal amount of inputs to produce

the optimal mix of outputs given the production technology and the prices it faces.

A firm, or in our case a country, can lie on or within the frontier, and the distance

between actual output and the frontier output represents technical inefficiency. The

technique assumes that given inputs, a maximum attainable output exists. The

4 Also the revenues considered in other studies as a proxy for output face the same problem.

10

country’s production lies on the frontier, if it uses the inputs efficiently, or within the

frontier, if it uses the inputs inefficiently. The distance between the frontier and the

actual production point measures technical inefficiency. Over time, a country’s

performance relative to the frontier includes two factors. First, a country can become

more efficient, and get closer to the frontier. Second, the frontier itself can shift over

time. Frontier shifts reflect purely technological factors. In addition, a country can move

along the frontier by changing inputs. Hence, output growth can be thought of in terms

of three components; efficiency change, technological change, and input change. A

simple growth accounting analysis that only looks at input quantity and an exogenous

technology does not take into account the country’s efficiency in using its inputs and

the available technology. This is especially true for developing countries where serious

institutional obstacles can impede the production process and cause a misuse not only

of capital and labour, but also of the imported technology. A stochastic frontier analysis

evaluates these effects as technical inefficiency. By separating technical inefficiency

from other variables, a stochastic frontier analysis decomposes TFP into factors that

are external to the country (technological change) and factors that are internal to the

country (its ability to absorb and use the available inputs).

We allow for efficiency and productivity as important determinant of cross-country

growth performance. Thus, we focus on several aspects that are most relevant to

efficiency: openness to trade, the development of infrastructures, the level of literacy,

the importance of the tourism sector, and the level of communication technologies. In

this way the paper is able to catch productivity differences across countries on a

comparable basis.

3.The model and the methodology applied

“Attempts to characterize tourism market supply have been limited due to a general

lack of product definition and explicit corporation of external characteristics critical to

producing tourism output. Furthermore, there are important natural-resource-based

public goods that tourism uses in its production that defy empirical analysis due to their

non-priced and common-pool characteristics.”(Marcouiller 1998).

This paper uses a different approach to Marcouiller (1998) and explicitly models a

tourism quasi-production function that includes capital and labour as inputs like in a

generic production function. But in this case the modelling of tourism supply comprises

two parts: firstly the stochastic production frontier (the estimates of productivity from the

production function); and secondly the technical efficiency model, that examines the

determinants of the inefficiency.

3.1 The Stochastic Production Frontier

The basics of the stochastic production frontier define:

it it itY X β ε= + (1.1)

where

it it itu vε = + (1.2)

Parametric estimation assumes the same input trade-off and returns to scale for all

firms/countries. Stochastic frontiers make explicit distributional assumptions on the

unobserved productivity and this allows avoiding the endogeneity problems that

consistent estimation of the input parameters faces (Van Biesebroeck 2006). The

model (Aigner, Lovell and Schmidt, 1977) is a pooled frontier model, in which the error

term is divided into two components: a normally distributed error , capturing general

measurements errors and heterogeneity; and a half-normal random term

itv

5 ,

representing the technical inefficiency of firm i at time t relative to the best-practice

production frontier ( ) as a one-sided non negative disturbance, where ~

itu

itvlnitu = − A

11

5 The most common distribution assumption about the inefficiency term consists of half –normal, exponential, and truncated distribution.

iid and ~ iid itu ( 20, uN )σ+ 6, where ( )20, uN σ+ stands for the positive part of

a normal distribution. Both error components are assumed to be uncorrelated with each

other and the regressors. This model is estimated by Maximum Likelihood and the

inefficiency component is estimated from the residuals it it itu vε = + by the conditional

expectation , proposed by Jondrow et al. (1982).

An advantage of stochastic frontiers is that it allows to generalize the production

function in order to permit for more sophisticated specifications. A disadvantage of this

methodology is that it does not allow determining if the observed performance of a

particular observation is due to inefficiency or to random variation in the frontier. It just

estimates the mean inefficiency over the sample.

We selected two estimable functional forms with acceptable theoretical and application

properties: the Transcendental Logarithmic and the Cobb-Douglas Production

functions, which substantially differ because the translog is quadratic in the logarithms

of its inputs and does not require a priori homotheticity and separability assumptions.

The factor shares of output are independent of total output if the production function is

homothetic, and the elasticities of substitution are constant and equal for any pair of

inputs if the function is both homothetic and additive. We thus assume a quasi

production function of the following form:

0lnY a= +

lnY A

1 1

1 ln2

n n n

iji i j

β= =

+∑ ∑∑

lnL K

1lni jx x

=

1 2ln K

ln lni ia x (1.3)

using a Translog or

ln lnα β γ+ += +

itu it

12

6 An alternative interpretation of the term is that the firm-specific production function lies u below the most productive observation in the sample (Van Biesebroeck 2006).

13

using a Cobb-Douglas

Where:

• Y is the output in form of Tourism arrivals. This variable represents the international

inbound tourists who travel to a country outside their place of residence for a period

not exceeding 12 months and whose main purpose in visiting is other than a

remunerated activity in the country visited. The source is: The World Bank, World

Development Indicators. Tourist output is evaluated here by a non-financial measure

as in Cracolici, Nijkamp, and Rietveld (2008).

The Xs are the inputs used.

• x1 = L is the labour utilized: Travel & Tourism Total Employment (in thousands) that

shows the total number of people employed in the Travel & Tourism industry, including

those jobs associated with Travel & Tourism, such as airline caterers, laundry services,

food suppliers, wholesalers, accounting firms, etc, government agencies,

manufacturing and construction of capital goods and exported goods used in Travel &

Tourism and supplied commodities, including steel producers, lumber, oil production,

etc. Source: World Travel & Tourism Council. α is the elasticity of output to labour.

• x2 =K2 stands for public capital, Government Expenditure on Travel & Tourism (US$

Million) that consists of two main components: collective and individual expenditures.

The collective government expenditures include expenditures by government agencies

and departments associated with the Travel & Tourism industry made on behalf of the

community at large. This includes tourism promotion, aviation administration, security

services and sanitation services. Individual government expenditures include cultural

institutions (art museums), national parks and customs and immigration on behalf of

individual visitors. Source: World Travel & Tourism Council. γ is the elasticity of output

to public capital.

• x3 =K1 stands for private capital that has been derived from the constraint

14

2 1K K K= + (1.4)

And K is the Capital Investment in Travel & Tourism (US$ Million) which includes the

value of all investments made by Travel & Tourism providers (the private sector) and

government agencies (the public sector) to provide facilities, equipment and

infrastructure for visitors. Source: World Travel & Tourism Council. β is the elasticity of

output to private capital.

These functional forms allow for either increasing, or constant, or decreasing returns to

scale7, just the isoquant being convex, but cannot identify factor-bias in technological

change.

Different specifications of the model consider just the private capital, both the public

and private capital, and just the total capital. When considering just the private capital

the public sector will play a role in explaining the efficiency trough the infrastructure

frame.

We used both a cross section of 208 countries in the year 2004 and a panel with the

same countries for four years (1990, 1995, 2000, 2004). Unfortunately, we could use

the data over the four years just for the production function but not for the technical

efficiency model.

7 Variable returns to scale is more appropriate in the tourism sector, because, for example, when bed-nights are considered as the output, its changes are unpredictable. (Peypoch 2007)

3.2 The Technical Efficiency Model

The technical efficiency model is specified as

it it itU Z wϕ= + (1.5)

Where is a random variable, defined by the truncation of the normal distribution, itw itZ

is a vector of country specific characteristics, and represents technical inefficiency. itU

Thus our specification of the technical efficiency model becomes:

0

1 2 3 4

5 6

inf& & &

it it it it it it itU op hur tech Tshare w eaasiaeu casia la c me na nasasia subsa

δδ δ δ δδ δ

= + + + + ++ + + + ++ + +

+ (1.6)

We focus on several aspects that could be relevant in explaining technical efficiency

across countries:

1)inf: the role of the public sector through an infrastructure index (ranging from 0 to

100) that shows the level of infrastructure development, combining the Road Index, the

Sanitation Index and the Water Access Index. Source: The World Bank, World

Development Indicators.

2)op: the role of international trade: the openness to trade index (ranging from 0 to 100)

shows the level of a country's openness towards international trade and international

visitors. The Openness Index is an aggregate index combining the Visa Index, Tourism

Openness Index, Trade Openness Index and Taxes on International Trade Index.

Source: World Travel & Tourism Council.

3)hur: some characteristics of the inputs used as the human resource index proxied by

using the Education Index obtained from the 2001 UNDP report. The education Index

consists of the adult literacy rate and the combined primary, secondary and tertiary

gross enrolment ratios. Source: The Human Development Report UNDP; or tse: the

15

16

tertiary school enrolment that refers to the Gross enrolment Ratio of tertiary school

calculated as a ratio of total enrolment in the tertiary education, regardless of age, to

the population of the age group that officially corresponds to the level of education

shown8. Source: The World Bank, World Development Indicators. The correlation

among these variables is very high, so It will be included just one of them in the

equation.

4)tech: the level of communication technology through a technology index (ranging

from 0 to 100) that combines the Internet Index, Telephone Index, Mobile Index and Hi-

Tech Index. Source: World Travel & Tourism Council.

5)Tshare: the ratio between the sum of international tourism receipts and tourism

expenditure to GDP9; it is a proxy for the importance of the tourism in the country

called the tourism impact index. It measures the direct economic effects of the

international tourism industry in the economy. Source: The World Bank, World

Development Indicators.

All these variables are presumed to have a negative impact on technical inefficiency.

And then we add geographical country group dummies, where eaasia stands for East

Asia and Pacific, eu & casia for Europe and Central Asia, la & c for Latin America and

Caribbean, me & na for Middle east and North Africa, na for North America, sasia for

South Asia and subsa for Sub Saharan Africa10.

8 Tertiary education, whether or not to an advanced research qualification, normally requires, as a minimum condition of admission, the successful completion of education at the secondary level. 9 International tourism receipts are the expenditures made by international inbound visitors, including payments to national carriers for international transport as well as any other pre-payment made for goods or services received in the destination country. Receipts may include expenditures made by same-day visitors, except in cases where these are so important as to justify a separate classification. International tourism expenditures are the expenditures made by international outbound visitors in other countries, including payments to foreign carriers for international transport. These may include expenditures by residents travelling abroad as same-day visitors, except in cases where these are so important as to justify a separate classification. 10 All the regressors in the equation (1.6) are proxies of the level of countries development, thus it was not necessary to add dummies referring to different stages of development.

17

4.Results

The theoretical models previously described have been estimated and on average the

results confirm the implicit hypothesis and reveal differences in the role played by the

two kinds of capital and the labour variables. The coefficients of the panel confirm the

prediction of the cross section. Owing to problems of multicollinearity present in the

data, the Cobb–Douglas functional form proved to be more appropriate with respect to

the Translog function.

Looking at the coefficients of our regressions, the most relevant variable to output is

the labour one confirming that tourism is a labour intensive sector. Between the two

kinds of capital the more important one is the private capital variable, while the public

capital is nor important or significant to output, while the second model suggests that

public investments and subsidies concerning tourism affect more the framework in

which tourist operators work.

The first model allows us to test for the elasticities of labour, public and private capital

to output11. See Table 1 for the results from the cross section and table 2 for the panel.

When separating total capital into private and public, When separating total capital into

private and public, the elasticity of output to public capital is not econometrically

significant and sometimes even with the opposite expected sign (negative); and the

elasticity of output to labour is on average 1.25: that is, a 1% increase in labour would

lead to an increase of international arrivals of 1.25 percentage points. Adding a dummy

relative to the level of income the capital is not significant, while the dummy is

significant, and being a rich country affect positively the output, while being a poor

country affect it negatively at 1% significance level. For the subset of low income

countries the elasticity of output to government expenditure is 0.74.

11 As far as the capital variable is concerned, even though the theoretical model includes separately both private and public capital, in the empirics attempts have been made also using a measure representing the total capital (private plus public) and using just the private capital variable. The results take into accounts also the results from these empirics.

18

As far as the technical efficiency model is concerned, the results suggest that the

human resource index, the level of communication technology, the country openness to

international trade all significantly contribute with the expected negative sign to

inefficiency. The variable regarding the importance of the tourism sector for the

economy of a country and the variable about the infrastructure level are never

significant. The country groups dummies reveal that there are some differences among

country groups: there is a proclivity in East Asia, Middle East, Europe and Central Asia,

and North America to affect positively the level of efficiency. See table 3.

5.Conclusion

Because of the growing importance of tourism, it has become fundamental to deepen

the empirical research on destination tourist performance and destination

competitiveness. In the service industry productivity definition and its measurement

becomes a difficult task due to different reasons. It is difficult to define and measure

“input and “output” in some services cases. Of course depending on the measure used

we will have different results. But the state of the art of literature in this field still rely on

a limited set of studies based on qualitative data.

Moreover, there is a lack of innovation in tourism. The tourism industry also suffers

from below-average labour productivity (OECD 2006). And tourism growth affects

economic growth in low and middle income countries, but not in rich countries.

If an economic unit (in our case a country) is operating in a in a competitive market,

managers and workers may feel pressure to work more efficiently and vice versa.

Obtaining a high efficiency estimate means being productive and this is consistent with

a competitive market . And because the hotel industry is not a highly regulated sector

(thus close to a competitive environment), and this industry is composed by many firms

that offer differentiated services, we should expect high efficiency. The quality of

managerial efficiency of the hotel industry could differ due to different causes (Hwang

19

and Chang 2002): 1) Market conditions: results demonstrate that leisure hotels are

better managed than urban hotels. Thus developing a low-season market strategy

becomes important. 2) Sources of customers: hotels in Taiwan whose customers are

mainly foreigners are more efficient than those only servicing local customers. 3)

Management style: “Compared to the independently operated local hotels, international

franchise-chains hotels have sounder reputation, better brand image, internet

marketing, efficient reservation system and economy of scale” (Hwang and Chang

2002).

This is an explorative econometric paper, that used a stochastic frontier approach to

estimate productivity in tourism all over the world in the years 1990, 1995, 2000 and

2004, and a technical efficiency model to analyze the determinants of efficiency across

countries.

As in Cracolici, Nijkamp and Rietveld (2008) the study found a proper way of dealing

with different indicators on the supply side in order to explain how countries transform

their resources into tourism flow.

The pioneer contribution of this study is threefold: 1) differently to previous studies of

this kind (Hwang and Chang 2002, Barros 2004, Cracolici et al. 2008) it analyzes

tourism supply and productivity at a worldwide level; 2) it proposes a way to model

tourism supply by estimating productivity from a quasi-production function of tourism; 3)

it analyzes if the differential of tourism productivity across countries could be due to

some structural characteristics of the countries themselves.

As far as the estimation of the production function is concerned, and trusting the

reliability of the data, the paper demonstrate that private capital and labour result to be

much more influential than public capital on the number of arrivals.

The state is responsible for ensuring a peaceful society, individual freedom, the rule of

law, security in terms of health, reliable public services and son on; moreover it

provides the legislative and regulatory framework for tourism. “Government should not

20

subsidise outright innovation, so as to avoid opportunism and rent-seeking behaviour.

Government’s role should be limited to that of a facilitator, coach or “incubation”

partner, passing prototypes to the private sector at the end of the innovation process

(Weiermair OECD2006). That is to say that one of the prescription of this study is to

recommend the public sector not to subsidizing hotel industry directly but supporting

and investing on the environment and legal framework in which the tourism firms

compete.

The paper asks the following questions: “How influential is productivity (in terms of

technical efficiency) in determining the number of arrivals? Are systematic differences

in efficiency among countries?”. The results demonstrate that efficiency contributes to

the number of arrivals and there are systematic differences among countries.

Although there are many factors that influence the levels and changes in tourist

supply, the level of literacy, the role exercised by the international trade trough the

openness to foreign markets, and the level of communication technologies, as far as

the results of this paper are concerned, are very important in explaining different level

of productivity across countries. That is to say that the more these variables are taken

into consideration in a country, the more productive the tourism industry is in that

country, the more these countries are able to compete in the world market.

Further developments of this research, depending on the availability of data will test if

countries with a higher productivity in tourism grow faster.

21

References

Aghion, P. and Howitt, P. (1998), “Endougenous Economic Growth”. Cambridge, MA:

the MIT Press.

Aigner, D, Lovell, C.A.K., and Schmidt, P. (1977), "Formulation and Estimation of

Stochastic Frontier Production Function Models", Journal of Econometrics, 6, pp.21-37.

Anderson. R.I., Fish, M., Xia Y., and Michello, F. (1999) “Measuring efficiency in the

hotel industry: a stochastic frontier examination” International Journal of Hospitality

Management, 18: (1999) 45-57

Anderson, R.I., Fok, R., Scott, J. “Hotel Industry Efficiency: an Advanced Linear

Programming Examination,” American Business Review, 18,1 (2000), 40-8

Banker, R.D. (1993), ‘Maximum likelihood, consistency and data envelopment

analysis’, Management

Science, Vol 39, No 10, pp 1265–1273.

Barro, R. J. (1990) “Government Spending in a Simple Model of Endogenous Growth”

The Journal of Political Economy 98(5(2)): S103-125.

Barro, R. J. (1991) “Economic growth in a Cross Section of Countries” The Quarterly

Journal of Economics 106(2):406-443.

Barros, C. P. “ A Stochastic Cost Frontier in the Portuguese Hotel Industry.” (2004)

Tourism Economics, pp. 177-192(16)

Barros, C.P., and Athanassiou, E. (2004), ‘Efficiency in European seaports with DEA:

evidence from

Greece and Portugal’, Maritime Economics & Logistics, Vol 6, No 2, pp 122–140.

Barros, C.P., and Alves, P. (2004), ‘Productivity in the tourism industry’, International

Advances in

Economic Research, Vol 10, No 3, pp 215–225.

22

Bell, R. A., Morey, R. C. “Increasing the Efficiency of Corporate Travel Management

through Macro benchmarking”, Journal of Travel Research, 33, 3, 1995, 11-20.

Berger, A.N., 1991. The proÞt-concentration relationship in banking. Finance and

Economics Discussion

Series (FEDS) vol. 176. Board of Governors of the Federal Reserve System.

Berger, A. N., 1993. “Distribution-free estimates of efficiency in the U:S: banking

industry and tests of the standard distribution assumptions.” Journal of Productivity

Analysis 4(3), 211-225.

Berger, A. N., Humphrey, D. B., 1991. “The dominance of inefficiencies over scale and

product mix economies in banking.” Journal of Monetary Economics 28(1), 117-148.

Berger, A.N., Humphrey, D.B., 1992a. Megamergers in banking and the use of cost

e¦ciency as an antitrust

defense. Antitrust Bulletin 37(3), 541Ð600.

Berger, A. N., Humphrey, D. B., 1992b. “Measurement and efficiency issues in

commercial banking. In: Griliches, Z. (Ed), M3easurement issues in Service Sectors,

NBER. Chicago Press, University of Chicago. IL.

Bessent, A.M., and Bessent, E.W. (1980), ‘Determining the comparative efficiency of

schools through

data envelopment analysis’, Educational Administration Quarterly, Vol 16, pp 57–75.

Biehl, D., Determinants of Regional Disparities and the Role of Public Finance, Public

Finance, vol. 35, 1980, pp. 44-71.

Blomstrom, M., Lipsey, R. and Zejan, M. (1996) “Is fixed investment the key to

economic growth?”, The Quarterly Journal of Economics, vol. 111, pp. 269-76.

Bradford DeLong, J. and Summers, L. (1991) “Equipment Investment and Economic

Growth”, The Quarterly Journal of Economics vol. 106:2, pp.445-502.

23

Brown, J.R., and Ragsdale, C.T. (2002), ‘The competitive market efficiency of hotel

brands: an

application of data envelopment analysis’, Journal of Hospitality & Tourism Research,

Vol 26, No 4, pp 332–360

Chiang, W.E., Tsai, M.H., and Wang, L.S.M. (2004), ‘A DEA evaluation of Taipei

hotels’, Annalsof Tourism Research, Vol 31, No 3, pp 712–715.

Cooper, W.W., Seiford, L.M., and Tone, K. (2000), Data Envelopment Analysis, Kluwer,

Boston, MA.

Cracolici, M. F., Nijkamp, P. and Rietveld. P. (2008) “Assesment of tourism

competitiveness by analyzing destination efficiency” Tourism Economics 14(2), 325-

342.

Eaton, J. and S. Kortum (1995) "Engines of Growth: Domestic and Foreign Sources of

Innovation" NBER Working Papers 5207.

Eaton, J. and S. Kortum (1996) "Measuring Technology Diffusion and the International

Sources of Growth" Eastern Economic Journal, Eastern Economic Association, vol.

22(4), pages 401-410, Fall.

Eugenio-Martín, J. L. Noelia Martín Morales and Riccardo Scarpa “Tourism and

Economic Growth in Latin American Countries: A Panel Data Approach” nota di lavoro

26.2004 FEEM.

Fare, R., S. Grosskopf, et al. (1994) “Productivity Growth, Technical Progress, and

Efficiency Change in Industrialized Countries”. American Economic Review 84(1)

:66-83.

Farrell, M.J. (1957) “The Measurement of Productive Efficiency.” Journal of the Royal

Statistical Society 120(3):253-290.

24

Feenstra Robert C., (1996) "Trade and uneven growth," Journal of Development

Economics, Elsevier, vol. 49(1), pages 229-256, April.

Grossman, G. M and E. Helpman (1991) Innovation and Growth in the Global

Economy, Cambridge, Mass., MIT Press.

International Labour Organization (2001). Human resources development, employment

and globalization in the hotel, catering and tourism sector, Geneva.

Iparraguirre D’elia. J. L. (2006) “The Five drivers of Productivity. How much does each

one contribute? Casual Analysis of Regional Labour Productivity in the UK.” ERINI

Monographs No. 14.

Jondrow J., Lovell C.A.K., Materov I.S., Schmidt P. (1982), On the Estimation of

Technical Inefficiency in the Stochastic Frontier Production Function Model, Journal

Econometrics.

Jorgenson, D. (2005) “Accounting for Growth in the Information Age”. In Philippe

Aghion and Steven Durlauf, eds., Handbook of Economic Growth, Vol. 1A, Amsterdam,

North-Holland, pp743-815

Hwang, S:N:, Chang, T:Y:, “Using data envelopment analysis to measure hotel

managerial efficiency change in Taiwan” Tourism Management 24 (2003): 357-369

Lanza A. and F. Pigliaru (1999) “Why are tourism countries small and fast growing?” A.

Fossati and G. Panella (eds), Tourism and Sustainable Economic Development,

Dordrecht: Kluwer Academic Publisher, 57-69.

Lozano. J., Gomez C., and J. Rey-Maquiera (2005) “An analysis of the evolution of

tourism destinations from the point of view of the economic growth theory” nota di

lavoro 146.2005 FEEM.

Lucas, R. E., Jr (1988), “On the mechanics of Economic Development” Journal of

Monetary Economics 22(1):3-42.

25

Mankiw, N., Romer, P. and Weil, D. (1992) “A contribution to the empirics of Economic

Growth”, The Quarterly Journal of Economics, vol. 107(2), pp. 407-37.

Marcouiller, D.W. (1998) Environmental resources as latent primary factors of

production in tourism: The case of forest-based commercial recreation. Tourism

Economics 4, 2: 131-145.

Morey, R.C., Dittman, D.A., 1995. “Evaluating a hotel GM’s performance: a case study

in benchmarking.” Cornell Hotel restaurant & Administration Quarterly 36(5), 30-35.

Nelson, R. & Phelps, E. (1966) “Investment in humans, technological diffusion, and

economic growth”, in American Economic Review, vol. 56, n. ½, pp. 65-75.

Neves Sequeira, Tiago and Carla Campos, 2005. "International Tourism and Economic

Growth: a Panel Data Approach," Working Papers 2005.141, Fondazione Eni Enrico

Mattei

Nickell, S and Van Reenen, J. (2001) “Technological Innovation and Economic

Performance in the United

Kingdom” Discussion paper No CEPDP0488. LSE

Nijkamp, P., A Multidimensional Analysis of Regional Infrastructure and Economic

Development, Structural Economic Analysis and Planning in Time and Space, (A.

Andersson and T. Puu, eds.), North-Holland Publ. Co., Amsterdam, 1982a, pp. 1-24.

Peypoch, N., (2007) “On Measuring Tourism Productivity”, Asia Pacific Journal of

Tourism Research, vol.12(3).

Pissarides Christopher A, (1997) "Learning by Trading and the Returns to Human

Capital in Developing Countries," World Bank Economic Review, Oxford University

Press, vol. 11(1), pages 17-32, January.

Rivera-Batiz, Luis A. & Romer, Paul M., 1991. "International trade with endogenous

technological change," European Economic Review, Elsevier, vol. 35(4), pages 971-

1001, May.

26

Romer, P.M. (1986) “Increasing Returns and Long-Run Growth” Journal of Political

Economy 94: 1002-1036.

Romer, P.M. (1990) “Endogenous Technological Change” Journal of Political Economy

98(5): pp. S71-S102 Part 2.

OECD (2006) Innovation and Growth in Tourism.

Schumpeter, J. (1934) “Theory of Economic Development” Harvard University Press,

Cambridge, MA.

UNDP The Human Development Report .

Van Biesebroeck 2006 “The sensitivity of Productivity Estimates: Revising Three

Important Debates” forthcoming in Journal of Business and Economic Statistics.

World Bank, World Development Indicators.

WTTC’s Competitiveness Monitor.

27

Table 1 Dep Var: lnY with separate K in the cross section

regressors coefficients (standard errors)

privk .09431161 ** .4377597

public k -.2790074 .3673533

employ 1.257269 *** .3214064

Log likelihood=-

406.64526

Wald

chi2(4)=1239.34

28

Table 2 Dep Var: lnY with separate K in the panel

regressors coefficients (standard errors)

privk 1.030949 *** .2182554

public k -.7492693 *** .1803824

employ 1.05286 *** .1485589

R-squared =

0.9083

F( 4, 567) =

1404.24

29

Table 3 Dep Var: technical inefficiency

regressors coefficients (standard errors)

Technology index -.0977836 *** .0365435

Openness index -.0020147 * .0131642

Infrastructure index -.0021535 .0010038

Tourism impact index -.0376333 .043587

Human resources

index .0076231 *** .0181172

easteasia -.1070559 *** .0356419

europeandcentralasia -.0784788 *** .0378086

latinamerica -.0565697 .031821

middleeast -.0768763 *** .0371436

northamerica -.1536005 *** .06436

southasia -.0709636 .0388611

Log likelihood = -

138.86011

Wald chi2(3) =

6.27

30